Date: Tue, 11 Mar 1997 18:29:54 +0100 From: Eivind Eklund <eivind@dimaga.com> To: =?KOI8-R?B?4c7E0sXKIP7F0s7P1w==?= <ache@nagual.ru> Cc: Bruce Evans <bde@zeta.org.au>, brian@awfulhak.demon.co.uk, brian@utell.co.uk, current@freebsd.org Subject: Re: ppp Message-ID: <3.0.32.19970311182952.00ca33a0@dimaga.com>
next in thread | raw e-mail | index | archive | help
At 06:41 PM 3/11/97 +0300, =?KOI8-R?B?4c7E0sXKIP7F0s7P1w==?= wrote: >On Wed, 12 Mar 1997, Bruce Evans wrote: > >> All systems. The kernel doesn't count signals, so if ppp isn't run for >> more than 1.2 seconds then it will miss some alarms. However, this probably >> isn't important since it will miss much more than alarms :-). There may be >> problems with closer-spaced alarms if ppp wants to count them. > >The big difference here is pended (delayed) and not pended signals >handling. PPP code was written in assumptions that SIGALRM reaction >is not delayed, so if we trust developer, he choose timeout + handler >action in the way that signals not missed. That would surprise me. Every other 'weird' condition (like malloc returning NULL, or lines not being the apropriate length) seems to be assumed to never happen - why should we assume the signal handling was robust when nothing else is? (I'll try to finish off the patches to those problems and send to Brian Real Soon Now.) Eivind Eklund perhaps@yes.no http://maybe.yes.no/perhaps/ eivind@freebsd.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.32.19970311182952.00ca33a0>